Agency

Definition

Agency in the context of AI-mediated literacy practices refers to the student's capacity to make intentional, informed decisions about how to engage with AI tools. It encompasses:

Control over the direction and scope of AI interactions
Choice about when and how to use AI assistance
Responsibility for evaluating and validating AI outputs
Authorship in determining what appears in final work

Agency is not about avoiding AI, but about maintaining deliberate control over the collaboration.

Theoretical Foundation

This conceptualization of agency draws from:

Sociocultural theory - Learners as active constructors of knowledge (Vygotsky, 1978)
Critical literacy - Agency as resistance to passive consumption (Freire, 1970)
Human-computer interaction - User agency in automated systems
Mollick & Mollick (2023) - Students must remain the "human in the loop"

Dimensions of Agency in AI-Literacy Practice

1. Agency over Knowledge Base (Co-Constructing AI Boundaries Framework Component - Inputs)

Selecting sources that challenge or expand AI's context
Curating materials that reflect critical perspectives
Framing problems in ways that resist AI's defaults

2. Agency over Cognitive Task (Co-Constructing AI Boundaries Framework Component - Prompts)

Crafting complex prompts that demand thinking
Constraining AI to prevent over-delegation
Asking for critique, synthesis, or comparison (not just summary)

3. Agency over Output Evaluation (Co-Constructing AI Boundaries Framework Component - Outputs)

Critically assessing quality and bias
Recognizing hallucinations or errors
Deciding what merits trust

4. Agency over Final Product (Co-Constructing AI Boundaries Framework Component - Integration)

Modifying, transforming, or rejecting AI text
Maintaining ownership of intellectual work
Ensuring human voice remains central

5. Metacognitive Agency (Co-Constructing AI Boundaries Framework Component - Reflection)

Articulating reasons for decisions
Reflecting on ethical dimensions
Positioning self as knowledge authority

Indicators of High vs. Low Agency

Indicator	High Agency	Low Agency
Prompts	Complex, constraining, directive	Generic, open-ended
Integration	Heavy modification/transformation	Copy-paste verbatim
Evaluation	Critical assessment, fact-checking	Uncritical acceptance
Reflection	Explicit boundary articulation	Minimal metacognition
Stance	"I used AI as a tool"	"AI did this for me"

Agency vs. Delegation

Important distinction:

Agency = Intentional, bounded delegation with oversight
Over-delegation = Relinquishing responsibility and critical evaluation

Students with high agency may delegate extensively while maintaining control through prompt design, evaluation, and transformation.

Relationship to Other Concepts

Boundary-work - Agency is enacted through boundary-setting actions
Epistemic Stance - Agency reflects how students position themselves as knowers

Evidence in This Study

[Add examples from student data showing high/low agency]

Example: High Agency in Prompts

"Act as a critical reviewer of this argument. Identify logical
fallacies and suggest counterarguments."

(Demands thinking, sets constraints, maintains oversight)

Example: Low Agency in Integration

[AI output inserted verbatim with no modification]

(No transformation, no voice, minimal authorship)

Pedagogical Implications

Fostering agency in AI-literacy requires:

Explicit instruction in prompt design
Structured reflection on decision-making
Emphasis on human responsibility
Models of critical evaluation
Practice with modification and transformation